Knowledge of the soil water characteristic (SWC) is crucial to understanding various soil-water related processes, plant growth, and to model gas and water flow. However, measurement of the SWC is expensive and time consuming, and until recently, the hyper-dry region of the SWC has been difficult to measure accurately. Based on recent attempts to derive the SWC from volumetric particle size fractions this study presents a new modified predictive model to estimate the volumetric water content from readily available soil data such as texture, bulk density and organic matter. Twenty-one source soils from Arizona and more than 200 Danish agricultural soils exhibiting a wide range of textures (clay 2-51%), organic carbon contents (OC 0-4%) and bulk densities (BD 1.0-1.75 g cm^-3) were used for model development. The new model estimates the representative pore size for each particle fraction and the corresponding water holding capacities. The capillary rise equation is employed to determine water retention at the cut-off matric potentials separating the volumetric pore fractions. The predicted volumetric water content – matric potential data pairs from saturation to -1.5 MPa are then used to parameterize the van Genuchten SWC model and a semi-log-linear model is fitted to the predicted points from -1.5 MPa to -100 MPa to obtain the continuous SWC. The model was tested for homogeneous and heterogeneous agricultural field soils and was able to capture the field scale variation of different soils and performed very well between near saturation and -100 MPa.